Method and Device for Detecting an Obstacle-Free Area

1. A method for forming a map of an obstacle-free area surrounding an image acquisition unit, which comprises the steps of: a) creating at least one image of a portion of an environment of the image acquisition unit by way of the image acquisition unit, and a digital image being created based on the at least one image; b) examining the digital image thus created for a presence of possible boundaries of the obstacle-free area, wherein a boundary indicator value is ascertained for individual pixels of the digital image, in dependence on a pixel-wise predefined image data of the digital image within an environment of a respective pixel, the boundary indicator value indicating a probability with which a map of a boundary of the obstacle-free area is disposed within a predefined pixel environment; c) predefining a coordinate parameterization having two parameters for the digital image; d) ascertaining a bounding curve of the obstacle-free area as a curve on which image points are disposed that, compared to image points in the environment thereof, have a high boundary indicator value; and e) establishing the map of the obstacle-free area by the bounding curve.

2. The method according to claim 1 , which further comprises selecting a polygonal chain that extends through a number of the pixels of the digital image as the bounding curve, a respective image point, in particular a pixel, through which the polygonal chain extends being selected for a number of predefined first parameters, and in particular for each image column of the digital image or for a number of image columns having a respective predefined x coordinate.

3. The method according to claim 1 , wherein for an ascertainment of the bounding curve, an optimization method is carried out in step d), in which, for individual bounding curves being considered, a respective measure value to be optimized is ascertained for individual image points on the respective bounding curve, each having predefined first parameters or x coordinates, the respective measure value being composed of the following partial measure values: the boundary indicator value; and a value of a second parameter or a value derived therefrom; and/or a difference between the second parameter in a particular point and the second parameter in a neighboring point located on a same bounding curve, the first parameter of which deviates by a predefined value, and which in particular is located in a neighboring column of the digital image, or a value derived from this difference.

4. The method according to claim 1 , wherein for the ascertainment of the bounding curve, an optimization method is carried out in step d), in which, for individual bounding curves being considered, in each case the image points on a particular bounding curve each having predefined first parameters or x coordinates, are selected which: compared to other image points have preferably high boundary indicator values; and/or have preferably high values of the second parameter; and/or have the second parameter which approximately corresponds to a second parameter of a particular neighboring point on the bounding curve, or has only a small difference compared to the second parameter, in particular so that a bounding curve results from a position of the image points whose change in a direction of the second parameter, is small.

5. The method according to claim 1 , which further comprises calculating the boundary indicator value in a pixel-wise manner by way of a neural network, the neural network receiving, as input values, individual brightness or color values within an environment around a respective pixel, the environment around the respective pixel being identically predefined for each pixel, in particular having a predefined relative position with respect to the respective pixel.

6. The method according to claim 5 , wherein the neural network is created using a plurality of predefined training images, a respective training bounding curve being predefined for each of the predefined training images, and the neural network being trained to supply a boundary indicator value that deviates from the boundary indicator values ascertained in a remaining image area, for a particular training image in the image points located in an area of the training bounding curve.

7. The method according to claim 1 , which further comprises ascertaining an actual size and a shape of a mapped obstacle-free area within the environment of the image acquisition unit, based on at least one of: a position of the obstacle-free area in the digital image; a relative position of the image acquisition unit compared to the environment; parameters of the image acquisition unit; or an assumption that the obstacle-free area acquired in the digital image is disposed on a predefined three-dimensional surface.

8. The method according to claim 1 , which further comprises acquiring the at least one image by way of the image acquisition unit being fixedly mounted on a person or on an animal.

9. The method according to claim 8 , wherein a position and an orientation of the image acquisition unit are monitored, and the image acquisition unit only creates images, or the steps b) to e) are only carried out, when the image acquisition unit is in a predefined position in which both an orientation of a horizontal image axis of the image acquisition unit and a viewing direction of the image acquisition unit are oriented approximately horizontally, or when a shoe containing the image acquisition unit is placed completely on a ground.

10. The method according to claim 8 , wherein the digital images are recorded by way of the image acquisition unit mounted on a shoe, a height of the image acquisition unit above ground being ascertained when the shoe is on the ground, and an actual size and shape of a mapped obstacle-free area being determined under a prerequisite that the image acquisition unit is located at an ascertained height.

11. The method according to claim 8 , wherein: in step b), in addition to the boundary indicator value, an obstacle indicator is also ascertained, which indicates a type of obstacle bounding the obstacle-free area; and the obstacle indicator is ascertained for individual image points on the bounding curve.

12. The method according to claim 1 , which further comprises: ascertaining a first distance of the bounding curve or of at least one point on the bounding curve of the obstacle-free area from the image acquisition unit; determining a second distance by distance measurement using ultrasound, an ultrasonic sensor used for measuring the second distance being disposed in an area of the image acquisition unit and being oriented in a same direction as the image acquisition unit, or having a detection area partially overlapping with an acquisition area of the image acquisition unit; and comparing the second distance and the first distance of the obstacle-free area to one another, and a further distance measurement value is created, using the first distance and the second distance, wherein, in an event of deviations between the first distance and the second distance, the lower distance of the first distance and the second distance is regarded as the further distance measurement value.

13. The method according to claim 1 , wherein: a distance value, corresponding to a size of the obstacle-free area, around a person is ascertained or predefined, and thereafter at least one of a plurality of actuator elements, which are disposed sequentially in a direction of a sensitive body region, is activated in at least one location of a body of the person, as a function of the distance value, the actuator elements being selected as a function of the distance value; and an actuator element in the sensitive body region being selected based on a smaller the distance value.

14. A device for forming a map of an obstacle-free area surrounding an image acquisition unit, the device comprising: said image acquisition unit; a processor connected downstream of said image acquisition unit; said image acquisition unit is configured to create at least one image of a portion of an environment of said image acquisition unit and to transmit the at least one image to said processor, said processor configured to: create a digital image based on the at least one image, and to examine the digital image thus created for a presence of possible boundaries of the obstacle-free area; ascertain a boundary indicator value for individual pixels of the digital image, as a function of pixel-wise predefined image data of the digital image within an environment of a particular pixel, said boundary indicator value indicating a probability with which a map of a boundary of the obstacle-free area is located within a predefined pixel environment; predefine a coordinate parameterization having two parameters for the digital image; ascertain a bounding curve of the obstacle-free area as a curve on which image points are located that, compared to image points in the environment thereof, have a high boundary indicator value; and establish the map of the obstacle-free area by the bounding curve.

15. The device according to claim 14 , wherein said processor is configured to: select a polygonal chain as the bounding curve, the polygonal chain extending through a number of the pixels of the digital image; and select a respective image point through which the polygonal chain extends for a number of predefined first parameters.

16. The device according to claim 14 , wherein said processor is configured to carry out an optimization method for an ascertainment of the bounding curve, said processor configured to ascertain, for individual bounding curves being considered, a respective measure value to be optimized for individual image points on a respective bounding curve, each having predefined first parameters or x coordinates, wherein the respective measure value is composed of the following partial measure values: the boundary indicator value; and a value of a second parameter or a value derived therefrom; and/or a difference between the second parameter in a particular point and the second parameter in a neighboring point located on a same bounding curve, a first parameter of which deviates by a predefined value, and is disposed in a neighboring column of the digital image, or a value derived from a difference.

17. The device according to claim 14 , wherein said processor is configured to carry out an optimization method for an ascertainment of the bounding curve, said processor configured, for individual bounding curves being considered, to select in each case the image points on a particular bounding curve each having predefined first parameters or x coordinates, which: compared to other image points have high boundary indicator values, and/or high values of the second parameter; and/or have the second parameter which approximately corresponds to the second parameter of a particular neighboring point on the bounding curve, or has only a small difference compared to the second parameter, so that, in particular, the bounding curve results from a position of the image points whose change in a direction of the second parameter is small.

18. The device according to claim 14 , wherein said processor is configured to: calculate the boundary indicator value in a pixel-wise manner by way of a neural network; and provide the neural network, as input values, with individual brightness or color values within an environment around a particular pixel.

19. The device according to claim 14 , wherein said processor is configured to: create a neural network using a plurality of predefined training images; and predefine a respective training bounding curve for each of the predefined training images, and to train the neural network to supply a boundary indicator value that deviates from boundary indicator values ascertained in a remaining image area, for a particular training image in image points disposed in an area of the respective training bounding curve.

20. The device according to claim 14 , wherein said processor is configured to ascertain an actual size and a shape of a mapped obstacle-free area within the environment of said image acquisition unit based on at least one of: a position of the obstacle-free area in the digital image; a relative position of said image acquisition unit compared to the environment; parameters of the image acquisition unit; or an assumption that the obstacle-free area acquired in the digital image is located on a predefined three-dimensional surface.

21. The device according to claim 14 , wherein said image acquisition unit is fixedly mounted on a person or on an animal.

22. The device according to claim 21 , wherein said processor is configured to monitor a position and an orientation of said image acquisition unit, said processor further configured to carry out further processing steps only when: said image acquisition unit is in a predefined position in which both an orientation of a horizontal image axis of said image acquisition unit and a viewing direction of said image acquisition unit are oriented approximately horizontally; or a shoe containing said image acquisition unit is placed completely on a ground; and/or said image acquisition unit is configured to create images only when: said image acquisition unit is in a predefined position in which both the orientation of the horizontal image axis of said image acquisition unit and the viewing direction of said image acquisition unit are oriented approximately horizontally; or the shoe containing said image acquisition unit is placed completely on the ground.

23. The device according to claim 22 , further comprising a shoe and said image acquisition unit is mounted in said shoe, said processor being configured to ascertain a height of said image acquisition unit above the ground when said shoe is on the ground, and to determine an actual size and shape of a mapped obstacle-free area under a prerequisite that said image acquisition unit is disposed at an ascertained height.

24. The device according to claim 14 , wherein said processor is configured to: ascertain, in addition to the boundary indicator value, an obstacle indicator, which indicates a type of obstacle bounding the obstacle-free area; and ascertain the obstacle indicator for individual ones of the image points on the bounding curve.

25. A device according to claim 14 , further comprising an ultrasonic sensor disposed in an area of said image acquisition unit, said ultrasonic sensor is oriented in a same direction as said image acquisition unit or has a detection area partially overlapping with an acquisition area of said image acquisition unit; wherein said processor is configured to: ascertain a first distance of the bounding curve or of at least one point on the bounding curve of the obstacle-free area from said image acquisition unit; determine a second distance by distance measurement by way of said ultrasonic sensor; and compare the second distance and the first distance of the obstacle-free area to one another, and to create a further distance measurement value, using the first distance and the second distance; and in an event of deviations between the first distance and the second distance, establish in each case a lower distance of the first distance and the second distance as the further distance measurement value.

26. The device according to claim 25 , further comprising a feedback unit, said feedback unit being configured to communicate data to said processing unit and to display an ascertained distance measurement value and/or an ascertained obstacle indicator value and/or an actual size and shape of a mapped obstacle-free area to the person.

27. The device according to claim 14 , further comprising a shoe containing a feedback unit having a plurality of actuator elements, said image acquisition unit and said processor, said shoe configured to: ascertain or predefine a distance value, corresponding to a size of the obstacle-free area, around a person; activate said feedback unit as a function of the distance value; activate, in said shoe of the person, at least one of a plurality of said actuator elements which are disposed sequentially in a direction of a sensitive body region; select said actuator elements as a function of the distance value, and to select an actuator element in said sensitive body region based on a smaller the distance value.